According to the New Goals for the Human Genome Project (1998-2003), 100,000-300,000 single nucleotide polymorphisms (SNPs) in the human genome will be discovered during the next five years. To utilize this valuable resource, robust and high-throughput SNP scoring technology needs to be developed. Although many high-throughput methods for post-PCR analysis are available, a primary prerequisite for SNP scoring, i.e., PCR amplification, has been a limiting step and presents a bottleneck for high-throughput SNP scoring. This is because with the conventional single-locus-based PCR, amplification of a large number of SNPs is very laborious and may not be affordable for many laboratories. As a step toward breaking this bottleneck, this application proposes to select 6,000 high-quality SNPs. Of these SNPs, 4,000 will be chosen from the databases containing SNPs with known heterozygosities in different human populations. The other 2,000 will be chosen after analyzing 6,000 SNPs without or with very limited information about their known heterozygosities in different human populations. As a by-product, the resulting information may serve as an important guideline for efficiently utilizing these polymorphisms. The 6,000 high-quality SNPs will be incorporated into 150 groups with 40 markers in each. Markers in each group will be robustly amplified in a multiplex way with a two-round PCR protocol. PCR products from the multiplex amplification will be analyzable by various genotyping approaches ranging from DNA-arrayed chips to commonly available fragmentation instruments. For the latter, restriction enzyme digestion is chosen for allelic sequence discrimination. Because greater than 60 percent SNPs can be analyzed by two restriction enzymes of 4-base recognition sites with or without single-base conversion during PCR, it is possible to have the markers in each group analyzable by a single enzyme and to analyze all selected SNPs by as few as two enzymes. The experimental conditions for the multiplex system will be worked out so that all groups will be analyzed by a single set of conditions with a sensitivity of detecting the target sequences in 1 ng genomic DNA.